The paper proposes a layer-wise curriculum learning strategy for compressing large language models (LLMs). By partitioning the model into layer segments and starting training with easier tasks before progressing to harder ones, the method accelerates convergence and stabilizes knowledge transfer from teacher to student models. Additional techniques such as feature caching with multi-threading improve GPU utilization, leading to state‑of‑the‑art compression results and over 50% reductions in memory usage and training time on BERT and GPT‑2, while outperforming other pruning methods on LLaMA‑family and Qwen models.
arXiv:2609.37988v1 Announce Type: new Abstract: As the context size of text processed with an LLM grows, the size of KV caches can outstrip the memory allocated for the original model weights. This i...
arXiv:2501.10375v3 Announce Type: replace-cross Abstract: Mixture-of-Experts (MoE) models, though highly effective for various machine learning tasks, face significant deployment challenges on memory...
arXiv:2607. 01844v1 Announce Type: cross Abstract: This paper showcases a memory-efficient training stack for Mixture-of-Experts (MoE) models.
arXiv:2608. 03796v1 Announce Type: cross Abstract: Small language models are often the only option for deployment under tight latency, cost, and on-premises constraints, but they are rarely trained from scratch: a compressed model is usually recovered through knowledge distillation (KD).
arXiv:2609.25537v1 Announce Type: new Abstract: Large language model (LLM) inference is constrained by the quadratic scaling of self-attention and the linear scaling of the KV cache, increasing laten...
MILO is a compression framework that reduces the key-value cache memory used in many-shot in-context learning by applying block-wise low-rank compression. It dynamically allocates rank budgets to blocks based on information entropy, preserving important information while aggressively compressing redundant parts. Experiments on Qwen2.5 models show up to a 50% reduction in KV cache memory and a 1.8× throughput improvement with negligible performance loss on classification and reasoning tasks.
arXiv:2506. 01260v2 Announce Type: replace Abstract: Scaling models has led to significant advancements in deep learning, but training these models in decentralized settings remains challenging due to communication bottlenecks.
The paper introduces a communication‑efficient method for adapting large language models on decentralized GPU meshes. It proposes an asynchronous two‑circuit system that uses fast compressed training with activation masking for pipeline‑parallel transfer and compressed data‑parallel synchronization, while a slower anchor circuit performs occasional unmasked passes. A spectral correction optimizer then denoises the masked gradients using these anchor priors, enabling high compression rates and achieving up to 40× throughput gains over internet‑grade connections while matching dense uncompressed performance.
Deep neural networks have witnessed remarkable advancements in recent years and have become integral to various applications. However, alongside these developments, training and deployment of neural network models on embedding and edge devices face significant challenges due to limited memory and computational resources.
arXiv:2606. 04063v1 Announce Type: cross Abstract: Deploying large language models (LLMs) is challenging due to their significant memory and computational requirements.
arXiv:2606. 10706v1 Announce Type: cross Abstract: Resource constraints increasingly determine what can be trained, fine-tuned, and deployed in large language models (LLMs), yet efficiency is often studied through isolated techniques rather than as an interacting system of limits.
arXiv:2504.04342v2 Announce Type: replace Abstract: Scaling up model parameters and training data consistently improves the performance of large language models (LLMs), but at the cost of rapidly gro...